Tcl Source Code

'\"
'\" Copyright (c) 1997 Sun Microsystems, Inc.
'\"
'\" See the file "license.terms" for information on usage and redistribution
'\" of this file, and for a DISCLAIMER OF ALL WARRANTIES.
'\"
.TH Tcl_UniCharIsAlpha 3 "8.1" Tcl "Tcl Library Procedures"
.so man.macros
.BS
.SH NAME
Tcl_UniCharIsAlnum, Tcl_UniCharIsAlpha, Tcl_UniCharIsControl, Tcl_UniCharIsDigit, Tcl_UniCharIsGraph, Tcl_UniCharIsLower, Tcl_UniCharIsPrint, Tcl_UniCharIsPunct, Tcl_UniCharIsSpace, Tcl_UniCharIsTransferable, Tcl_UniCharIsUpper, Tcl_UniCharIsWordChar \- routines for classification of Tcl_UniChar characters
.SH SYNOPSIS
.nf
\fB#include <tcl.h>\fR
.sp
int
\fBTcl_UniCharIsAlnum\fR(\fIch\fR)
.sp

int
\fBTcl_UniCharIsPunct\fR(\fIch\fR)
.sp
int
\fBTcl_UniCharIsSpace\fR(\fIch\fR)
.sp
int
\fBTcl_UniCharIsTransferable\fR(\fIch\fR)
.sp
int
\fBTcl_UniCharIsUpper\fR(\fIch\fR)
.sp
int
\fBTcl_UniCharIsWordChar\fR(\fIch\fR)
.SH ARGUMENTS
.AS int ch
.AP int ch in
The Unicode character to be examined.

.PP
\fBTcl_UniCharIsPrint\fR tests if the character is a Unicode print character.
.PP
\fBTcl_UniCharIsPunct\fR tests if the character is a Unicode punctuation character.
.PP
\fBTcl_UniCharIsSpace\fR tests if the character is a whitespace Unicode character.
.PP
\fBTcl_UniCharIsTransferable\fR tests if the character is a Unicode character, not being
a surrogate or noncharacter.
.PP
\fBTcl_UniCharIsUpper\fR tests if the character is an uppercase Unicode character.

.PP
\fBTcl_UniCharIsWordChar\fR tests if the character is alphanumeric or
a connector punctuation mark.

.SH KEYWORDS
unicode, classification

.IP \fBspace\fR 12
Any Unicode whitespace character, mongolian vowel separator
(U+180e), zero width space (U+200b), word joiner (U+2060) or
zero width no-break space (U+feff) (=BOM).
.IP \fBtrue\fR 12
Any of the forms allowed to \fBTcl_GetBoolean\fR where the value is
true.
.IP \fBtransferable\fR 12
Any Unicode character, except surrogates and noncharacters.
The purpose is to check, if a string may be transfered to an external program using an unicode encoding.
.IP \fBupper\fR 12
Any upper case alphabet character in the Unicode character set.
.IP \fBwideinteger\fR 12
Any of the valid forms for a wide integer in Tcl, with optional
surrounding whitespace.  In case of overflow in the value, 0 is
returned and the \fIvarname\fR will contain \-1.
.IP \fBwordchar\fR 12

declare 655 {
    const char *Tcl_UtfNext(const char *src)
}
declare 656 {
    const char *Tcl_UtfPrev(const char *src, const char *start)
}
declare 657 {
    int Tcl_UniCharIsTransferable(int ch)
}

# TIP 656
declare 658 {
    int Tcl_ExternalToUtfDStringEx(Tcl_Interp *interp, Tcl_Encoding encoding,
        const char *src, Tcl_Size srcLen, int flags, Tcl_DString *dsPtr,
        Tcl_Size *errorLocationPtr)

    Tcl_WideInt w;

    static const char *const isClasses[] = {
	"alnum",	"alpha",	"ascii",	"control",
	"boolean",	"dict",		"digit",	"double",
	"entier",	"false",	"graph",	"integer",
	"list",		"lower",	"print",	"punct",
	"space",	"transferable",	"true",		"upper",
	"wideinteger", "wordchar",	"xdigit",	NULL
    };
    enum isClassesEnum {
	STR_IS_ALNUM,	STR_IS_ALPHA,	STR_IS_ASCII,	STR_IS_CONTROL,
	STR_IS_BOOL,	STR_IS_DICT,	STR_IS_DIGIT,	STR_IS_DOUBLE,
	STR_IS_ENTIER,	STR_IS_FALSE,	STR_IS_GRAPH,	STR_IS_INT,
	STR_IS_LIST,	STR_IS_LOWER,	STR_IS_PRINT,	STR_IS_PUNCT,
	STR_IS_SPACE,	STR_IS_TRANSFERABLE,	STR_IS_TRUE,	STR_IS_UPPER,
	STR_IS_WIDE,	STR_IS_WORD,	STR_IS_XDIGIT
    };
    static const char *const isOptions[] = {
	"-strict", "-failindex", NULL
    };
    enum isOptionsEnum {
	OPT_STRICT, OPT_FAILIDX

	break;
    case STR_IS_SPACE:
	chcomp = Tcl_UniCharIsSpace;
	break;
    case STR_IS_UPPER:
	chcomp = Tcl_UniCharIsUpper;
	break;
    case STR_IS_TRANSFERABLE:
	chcomp = Tcl_UniCharIsTransferable;
	break;
    case STR_IS_WORD:
	chcomp = Tcl_UniCharIsWordChar;
	break;
    case STR_IS_XDIGIT:
	chcomp = UniCharIsHexDigit;
	break;

    DefineLineInformation;	/* TIP #280 */
    Tcl_Token *tokenPtr = TokenAfter(parsePtr->tokenPtr);
    static const char *const isClasses[] = {
	"alnum",	"alpha",	"ascii",	"control",
	"boolean",	"dict",		"digit",	"double",
	"entier",	"false",	"graph",	"integer",
	"list",		"lower",	"print",	"punct",
	"space",	"transferable",	"true",		"upper",
	"wideinteger", "wordchar",	"xdigit",	NULL
    };
    enum isClassesEnum {
	STR_IS_ALNUM,	STR_IS_ALPHA,	STR_IS_ASCII,	STR_IS_CONTROL,
	STR_IS_BOOL,	STR_IS_DICT,	STR_IS_DIGIT,	STR_IS_DOUBLE,
	STR_IS_ENTIER,	STR_IS_FALSE,	STR_IS_GRAPH,	STR_IS_INT,
	STR_IS_LIST,	STR_IS_LOWER,	STR_IS_PRINT,	STR_IS_PUNCT,
	STR_IS_SPACE,	STR_IS_TRANSFERABLE,	STR_IS_TRUE,	STR_IS_UPPER,
	STR_IS_WIDE,	STR_IS_WORD,	STR_IS_XDIGIT
    };
    int t, range, allowEmpty = 0, end;
    InstStringClassType strClassType;
    Tcl_Obj *isClass;

    if (parsePtr->numWords < 3 || parsePtr->numWords > 6) {

	goto compileStrClass;
    case STR_IS_PUNCT:
	strClassType = STR_CLASS_PUNCT;
	goto compileStrClass;
    case STR_IS_SPACE:
	strClassType = STR_CLASS_SPACE;
	goto compileStrClass;
    case STR_IS_TRANSFERABLE:
	strClassType = STR_CLASS_TRANSFERABLE;
	goto compileStrClass;
    case STR_IS_UPPER:
	strClassType = STR_CLASS_UPPER;
	goto compileStrClass;



    case STR_IS_WORD:
	strClassType = STR_CLASS_WORD;
	goto compileStrClass;
    case STR_IS_XDIGIT:
	strClassType = STR_CLASS_XDIGIT;
    compileStrClass:
	if (allowEmpty) {

    {"lower",	Tcl_UniCharIsLower},
    {"print",	Tcl_UniCharIsPrint},
    {"punct",	Tcl_UniCharIsPunct},
    {"space",	Tcl_UniCharIsSpace},
    {"upper",	Tcl_UniCharIsUpper},
    {"word",	Tcl_UniCharIsWordChar},
    {"xdigit",	UniCharIsHexDigit},
    {"transferable",	Tcl_UniCharIsTransferable},
    {"",	NULL}
};

/*
 *----------------------------------------------------------------------
 *
 * TclCompileSubstCmd --

    STR_CLASS_PUNCT,		/* Unicode punctuation characters. */
    STR_CLASS_SPACE,		/* Unicode space characters. */
    STR_CLASS_UPPER,		/* Unicode upper-case alphabet characters. */
    STR_CLASS_WORD,		/* Unicode word (alphabetic, digit, connector
				 * punctuation) characters. */
    STR_CLASS_XDIGIT,		/* Characters that can be used as digits in
				 * hexadecimal numbers ([0-9A-Fa-f]). */
    STR_CLASS_TRANSFERABLE	/* Transferable characters. */
} InstStringClassType;

typedef struct StringClassDesc {
    char name[16];		/* Name of the class. */
    int (*comparator)(int);	/* Function to test if a single unicode
				 * character is a member of the class. */
} StringClassDesc;

MODULE_SCOPE StringClassDesc const tclStringClassTable[];

/*

/* 654 */
EXTERN int		Tcl_UtfCharComplete(const char *src, Tcl_Size length);
/* 655 */
EXTERN const char *	Tcl_UtfNext(const char *src);
/* 656 */
EXTERN const char *	Tcl_UtfPrev(const char *src, const char *start);
/* 657 */
EXTERN int		Tcl_UniCharIsTransferable(int ch);
/* 658 */
EXTERN int		Tcl_ExternalToUtfDStringEx(Tcl_Interp *interp,
				Tcl_Encoding encoding, const char *src,
				Tcl_Size srcLen, int flags,
				Tcl_DString *dsPtr,
				Tcl_Size *errorLocationPtr);
/* 659 */

    void (*reserved650)(void);
    void (*reserved651)(void);
    void (*reserved652)(void);
    void (*reserved653)(void);
    int (*tcl_UtfCharComplete) (const char *src, Tcl_Size length); /* 654 */
    const char * (*tcl_UtfNext) (const char *src); /* 655 */
    const char * (*tcl_UtfPrev) (const char *src, const char *start); /* 656 */
    int (*tcl_UniCharIsTransferable) (int ch); /* 657 */
    int (*tcl_ExternalToUtfDStringEx) (Tcl_Interp *interp, Tcl_Encoding encoding, const char *src, Tcl_Size srcLen, int flags, Tcl_DString *dsPtr, Tcl_Size *errorLocationPtr); /* 658 */
    int (*tcl_UtfToExternalDStringEx) (Tcl_Interp *interp, Tcl_Encoding encoding, const char *src, Tcl_Size srcLen, int flags, Tcl_DString *dsPtr, Tcl_Size *errorLocationPtr); /* 659 */
    int (*tcl_AsyncMarkFromSignal) (Tcl_AsyncHandler async, int sigNumber); /* 660 */
    void (*reserved661)(void);
    void (*reserved662)(void);
    void (*reserved663)(void);
    void (*reserved664)(void);

/* Slot 653 is reserved */
#define Tcl_UtfCharComplete \
	(tclStubsPtr->tcl_UtfCharComplete) /* 654 */
#define Tcl_UtfNext \
	(tclStubsPtr->tcl_UtfNext) /* 655 */
#define Tcl_UtfPrev \
	(tclStubsPtr->tcl_UtfPrev) /* 656 */
#define Tcl_UniCharIsTransferable \
	(tclStubsPtr->tcl_UniCharIsTransferable) /* 657 */
#define Tcl_ExternalToUtfDStringEx \
	(tclStubsPtr->tcl_ExternalToUtfDStringEx) /* 658 */
#define Tcl_UtfToExternalDStringEx \
	(tclStubsPtr->tcl_UtfToExternalDStringEx) /* 659 */
#define Tcl_AsyncMarkFromSignal \
	(tclStubsPtr->tcl_AsyncMarkFromSignal) /* 660 */
/* Slot 661 is reserved */

    0, /* 650 */
    0, /* 651 */
    0, /* 652 */
    0, /* 653 */
    Tcl_UtfCharComplete, /* 654 */
    Tcl_UtfNext, /* 655 */
    Tcl_UtfPrev, /* 656 */
    Tcl_UniCharIsTransferable, /* 657 */
    Tcl_ExternalToUtfDStringEx, /* 658 */
    Tcl_UtfToExternalDStringEx, /* 659 */
    Tcl_AsyncMarkFromSignal, /* 660 */
    0, /* 661 */
    0, /* 662 */
    0, /* 663 */
    0, /* 664 */

    }
    return (GetCategory(ch) == UPPERCASE_LETTER);
}

/*
 *----------------------------------------------------------------------
 *
 * Tcl_UniCharIsTransferable --
 *
 *	Test if a character is a transferable Unicode character.
 *
 * Results:
 *	Returns non-zero if character belongs to the Unicode set.
 *
 *	Excluded are:
 *	  1) All characters > U+10FFFF
 *	  2) Surrogates U+D800 - U+DFFF
 *	  3) Last 2 characters of each plane, so U+??FFFE  and U+??FFFF
 *	  4) The characters in the range U+FDD0 - U+FDEF
 *
 * Side effects:
 *	None.
 *
 *----------------------------------------------------------------------
 */

int
Tcl_UniCharIsTransferable(
    int ch)			/* Unicode character to test. */
{
    return ((unsigned int)ch <= 0x10FFFF) && ((ch & 0xFFF800) != 0xD800)
	    && ((ch & 0xFFFE) != 0xFFFE) && ((unsigned int)(ch - 0xFDD0) >= 32);
}

/*

    list [catch {run {string is alpha -failin str}} msg] $msg
} {1 {wrong # args: should be "string is alpha ?-strict? ?-failindex var? str"}}
test string-6.4.$noComp {string is, too many args} {
    list [catch {run {string is alpha -failin var -strict str more}} msg] $msg
} {1 {wrong # args: should be "string is class ?-strict? ?-failindex var? str"}}
test string-6.5.$noComp {string is, class check} {
    list [catch {run {string is bogus str}} msg] $msg
} {1 {bad class "bogus": must be alnum, alpha, ascii, control, boolean, dict, digit, double, entier, false, graph, integer, list, lower, print, punct, space, transferable, true, upper, wideinteger, wordchar, or xdigit}}
test string-6.6.$noComp {string is, ambiguous class} {
    list [catch {run {string is al str}} msg] $msg
} {1 {ambiguous class "al": must be alnum, alpha, ascii, control, boolean, dict, digit, double, entier, false, graph, integer, list, lower, print, punct, space, transferable, true, upper, wideinteger, wordchar, or xdigit}}
test string-6.7.$noComp {string is alpha, all ok} {
    run {string is alpha -strict -failindex var abc}
} 1
test string-6.8.$noComp {string is, error in var} {
    list [run {string is alpha -failindex var abc5def}] $var
} {0 3}
test string-6.9.$noComp {string is, var shouldn't get set} {

} {0 87}
test string-6.130.1.$noComp {string is entier, false on bad octal} {
    list [run {string is entier -fail var 0o1234561123412345612345656234561234561234561234561234561234561234561234561234561234536963}] $var
} {0 87}
test string-6.131.$noComp {string is entier, false on bad hex} {
    list [run {string is entier -fail var 0X12345611234123456123456562345612345612345612345612345612345612345612345612345612345345XYZ}] $var
} {0 88}
test string-6.132.$noComp {string is transferable} {
    run {string is transferable \U10FFFD\uD7FF\uE000\uFDCF\uFDF0}
} 1
test string-6.133.$noComp {string is transferable, upper surrogate} {
    run {string is transferable \uD800}
} 0
test string-6.134.$noComp {string is transferable, lower surrogate} {
    run {string is transferable \uDFFF}
} 0
test string-6.135.$noComp {string is transferable, noncharacter} {
    run {string is transferable \uFFFE}
} 0
test string-6.136.$noComp {string is transferable, noncharacter} {
    run {string is transferable \uFFFF}
} 0
test string-6.137.$noComp {string is transferable, noncharacter} {
    run {string is transferable \uFDD0}
} 0
test string-6.138.$noComp {string is transferable, noncharacter} {
    run {string is transferable \uFDEF}
} 0
test string-6.139.$noComp {string is integer, bug [76ad7aeba3]} {
    run {string is integer 18446744073709551615}
} 1
test string-6.140.$noComp {string is integer, bug [76ad7aeba3]} {
    run {string is integer -18446744073709551615}
} 1